I just wanted to put 2-cent worth of my comment here about James Hooker's GE-Mazda Ribbon Filament Microscope Illuminator.  I noticed there two different glasses at the stem press area.  My guess they were attempting to make a good seal between the glass/metal using the glass/metal tubes prior pressing with the outside glass tube

See the stem press assembly detail at:

But that is pretty standard way, mainly when done by hand. Sealing first the small glass tubes/beads mean there is no hard glass around, so because all glass is melted/soft when you are attempting to heat up the heavy metal part (in order to wet it with the glass properly), there is nothing to crack due to thermal gradients.
Then with the second step sealing the main stem to these glass beads is way easier, as it is glass on glass, (the metal does not have to be fully heated anymore), just softening it a bit is enough to properly fuse together.

With mass production the two step process is to lengthy so expensive, so usually more time is spent to fine tune the process so it works still in one step (you spend the time once, then the machine repeats the process inherently very accurately). But on really large parts or with high-rel or long life products (where the robustness of the seal so as low as possible stress in the material is mandatory and the money budget for processing is higher) even the mass production uses two step process, it just gives better quality seals.

Well spotted!  This is all correct, but even today for mass produced lamps with hardglass stems it is very common to first sleeve the leadwires in small glass tubes, and then melt them into the stem.  However you cannot always see this becuase these days there is less difference in refractive index between the sleeving and stem glass.

The reason is because seals between borosilicate glass and tungsten only work when the tungsten has been pre-oxidised to the correct state, and when enough of this oxide is then dissolved into the surrounding glass.  The same is true of the grey coloured tungsten-molybdenum alloy seals used on GE/Tungsram/Philips lamps.  It is very difficult to form these seals correctly on the relatively high speed stem-making machines.  Therefore there are other offline machines that first oxidise the wires and then sleeve them in glass.  Usually this operation is done by the factories that weld up the leadwire assemblies, and the lamp factory then only has to fuse these wires into the stem assembly just the same as for any other lamp without having to take so many extra precautions.

Thanks for the explanation, James

James, a very nice lamp you have here. I also have a question about that. What is the difference between these and Lamp for scientific use (for example calibration lamps like Osram Wi series) wich also use a tungsten band. Are these special calibration lamps of higher quality or is there also another different?

Kind regards

Alex